parsing out one variable to many
Locked
parsing out one variable to many
 
|
I am collecting survey data. Respondents are asked to categorize their
comments into different buckets. They can choose as many as they like or none at all. If they choose the first category, the data puts a "1" in the field. If they choose the first and fifth category, the data reads "1,5." There are 15 possible choices. I'd like to create a variable called "strategy" for category 1 and have a "1" put in that variable if the respondent chose that bucket; a variable called "communication" for category 2 and a "1" put in that variable if the respondent chose that bucket and so on. sample data, variable named qID_9109: 1,2 3,6,7 2 5,10,12,14 Can anyone help here? ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: parsing out one variable to many
|
Nancy IN SPSS, easiest would be to have 15 variables, one for each response category. If you label each value (1st variable only) you can then use MULT RESPONSE to tabulate the responses. val lab v1 1 ‘Strategy’ 2 ‘ ~~~~’ etc etc mult resp category (v1 to v15 (1,15)) /freq category. Another way is to label each variable and use MULT RESPONSE in dichotomous mode, but to do this you first have to recode the values for all variables to 1. I’ve played safe and kept your originals, but recoded to a new variable recode v1 to v15 (1 thru 15 = 1)(else =0) into r1 to r15. var lab r1 ‘~~~’ /r2 ‘~~~’ ~~~ /r15 ‘~~~’. mult resp rcategory (r1 to r15 (1)) /freq rcategory. There are examples of this (illustrated by multiple response questions from the British Social Attitudes survey 1987 and the European Social Survey 2002) in sections 3.2 and 3.3 (pp 27-35) and the 3rd slide show of my 2006 Old Dog, Old Tricks presentation SPSS usage in major surveys. (See: http://surveyresearch.weebly.com/old-dog-old-tricks-using-spss-syntax-to-beat-the-mouse-trap.html) Get back to me if you don’t understand that or if you need further help. John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com Start page: www.surveyresearch.weebly.com/spss-without-tears.html -----Original Message----- I am collecting survey data. Respondents are asked to categorize their comments into different buckets. They can choose as many as they like or none at all. If they choose the first category, the data puts a "1" in the field. If they choose the first and fifth category, the data reads "1,5." There are 15 possible choices. I'd like to create a variable called "strategy" for category 1 and have a "1" put in that variable if the respondent chose that bucket; a variable called "communication" for category 2 and a "1" put in that variable if the respondent chose that bucket and so on. sample data, variable named qID_9109: 1,2 3,6,7 2 5,10,12,14 Can anyone help here? ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: parsing out one variable to many
Administrator
|
In reply to this post by Nancy Rusinak-2
FWIW: That's a silly way to enter data! Especially when you don't have a well thought out plan to turn it into something usable.
However, here is a mop up method. http://spssx-discussion.1045642.n5.nabble.com/Parsing-String-Name-Variable-tc5714037.html#a5714040
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
Re: parsing out one variable to many
|
In reply to this post by Nancy Rusinak-2
If you have not yet
entered the data.
You need to have as many variables as the maximum number that any respondent choose. For the data example you posted that would be 4 variables. maybe named qID9109c1 to qID9109c4 then you would deal with that via MULT RESP . This would be a multiple response set as opposed to a multiple dichotomy set. Are your data already entered so that you have string variables that look like the examples you posted? Art Kendall Social Research ConsultantsOn 4/22/2013 9:54 AM, Nancy Rusinak-2 [via SPSSX Discussion] wrote: I am collecting survey data. Respondents are asked to categorize their
Art Kendall
Social Research Consultants |
Re: parsing out one variable to many
|
|
In reply to this post by Nancy Rusinak-2
Here's a not so compact method. A key drawback is that when the loop exits after processing a case a warning is issued. I don't know whether there is a way to rewrite it to preclude the warning since you don't know how many responses are concatenated unless you scoot through each record and count the commas. Gene Maguin
String #s(a36). Vector y(15,f2.0). Compute #s=qID_9109. If (substr(#s,char.len(#s),1) ne ',') #s=concat(rtrim(qID_9109),','). Loop if (char.len(#s) ge 2). Compute #i=index(#s,','). If (#i eq 2) y(number(substr(#s,1,1),f1.0))=1. If (#i eq 3) y(number(substr(#s,1,2),f2.0))=1. if (char.len(#s) ge 4) #s=substr(#s,#i+1). End loop. -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Nancy Rusinak Sent: Monday, April 22, 2013 9:54 AM To: [hidden email] Subject: parsing out one variable to many I am collecting survey data. Respondents are asked to categorize their comments into different buckets. They can choose as many as they like or none at all. If they choose the first category, the data puts a "1" in the field. If they choose the first and fifth category, the data reads "1,5." There are 15 possible choices. I'd like to create a variable called "strategy" for category 1 and have a "1" put in that variable if the respondent chose that bucket; a variable called "communication" for category 2 and a "1" put in that variable if the respondent chose that bucket and so on. sample data, variable named qID_9109: 1,2 3,6,7 2 5,10,12,14 Can anyone help here? ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: parsing out one variable to many
Administrator
|
Gene, I think the problem was that your condition for exiting the loop was never becoming true, so it was only exiting the loop when it the loop number exceeded MXLOOPS. Given that we know 15 is the maximum number of loops (in this case, at least), if you LOOP just 15 times, it works with no warning messages. E.g.,
data list list / qID_9109(a36). begin data "1,2" "3,6,7" "2" "5,10,12,14" "1,2,3,4,5,6,7,8,9,10,11,12,13,14,15" end data. list. * Gene Maguin's code with some small modificiations. String #s(a36). Vector y(15,f2.0). Compute #s=qID_9109. If (substr(#s,char.len(#s),1) ne ',') #s=concat(rtrim(qID_9109),','). compute #k = 1. Loop #k = 1 to 15. Compute #i=index(#s,','). If (#i eq 2) y(number(substr(#s,1,1),f1.0))=1. If (#i eq 3) y(number(substr(#s,1,2),f2.0))=1. if (char.len(#s) ge 4) #s=substr(#s,#i+1). End loop. recode Y1 to Y15 (sysmis=0). LIST Y1 to Y15 qID_9109. OUTPUT: y1 y2 y3 y4 y5 y6 y7 y8 y9 y10 y11 y12 y13 y14 y15 qID_9109 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1,2 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 3,6,7 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 1 0 0 0 0 1 0 1 0 1 0 5,10,12,14 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 Number of cases read: 5 Number of cases listed: 5
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
Re: parsing out one variable to many
|
|
In reply to this post by John F Hall
Nancy I can see your problem, but other listers have suggested solutions to spread out the strings. I’m copying this back to the list so others can see it. How many cases do you have? Are there only up to three responses for each respondent? If so you need to spread the responses out on to three separate columns, which is what you’ve tried to do with your syntax. String manipulation is not my forte, I’m afraid. You can apply the same logic if there are more than three responses per case. I’m always aware that time = money and I don’t have time to learn about strings. Provided the data set is not too large, I would probably attempt something by hand such as copying the strings into two more columns and then editing them. I would also try copying the column for Q.66 into Word to edit into something that data list can read as three variables (even if the 2nd and 3rd are blank). Here’s my rough and ready solution: Copy your data into a txt file 0,1,14 2,6,7 10,11,15 3 0,4 8,10,12 Replace all single digits with leading 0 and commas by blanks (I did this by hand) 00 01 14 02 06 07 10 11 15 03 00 04 08 10 12 . . and save to nancy.txt on f: SPSS syntax: data list file 'f:\nancy.txt' rec 1 / x1 1-2 x2 4-5 x3 7-8. list /cases 6. freq var x1 x2 x3. mult resp groups q66 (x1 to x3 (0,15)) /freq q66. DATASET ACTIVATE DataSet9. data list file 'f:\nancy.txt' rec 1 / x1 1-2 x2 4-5 x3 7-8. Data List will read 1 records from f:\nancy.txt Variable Rec Start End Format x1 1 1 2 F2.0 x2 1 4 5 F2.0 x3 1 7 8 F2.0 list /cases 6. x1 x2 x3 0 1 14 2 6 7 10 11 15 3 . . 0 4 . 8 10 12 Number of cases read: 6 Number of cases listed: 6
Is this what you wanted? John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com Start page: www.surveyresearch.weebly.com/spss-without-tears.html From: Nancy Rusinak [mailto:[hidden email]] Hello Mr. Hall, Thank you for writing & attempting to assist me. I'm afraid I have never worked with MULT RESPONSE syntax in the past so it's all Greek to me at this point. I have worked a good bit with SUBSTR and am trying to create substrings using the "," as a delimiter. I'm almost there but, alas, not quite. Can you possibly help? Here's a sample of my data -variable is called Q66. It's string, 10 wide: 0,1,14 2,6,7 10,11,15 3 0,4 8,10,12 I only want 3 variables when I finish. For example, for the first row of data, variable 1 would be "0," variable 2 would be "1," and variable 3 would be "14." Here's my syntax, which I found online and have been trying to adapt for my purposes: STRING #TEMP(A50). COMPUTE #TEMP=Q66. VECTOR VAR(3, A2). LOOP #I=1 TO 3. COMPUTE #INDEX=INDEX(#TEMP, ","). COMPUTE VAR(#I)=SUBSTR(#TEMP,1, #INDEX-1). COMPUTE #TEMP=SUBSTR(#TEMP, #INDEX+1). END LOOP IF INDEX(#TEMP, ",")=0. EXECUTE. When I run this, the first two variables are correctly populated for the data strings that contain a "," but not for data strings that do not. Those are null. Also, the third variable is not populated for any of the data - just null down the column. SPSS output logs the following: "the third argument to SUBSTR (the length) is missing or otherwise invalid. The argument must be a non-negative integer. The result has been set to the null string." This indicates to me that the #INDEX-1 and #INDEX+1 are an issue but I do not know a workaround... Any help you could offer would be greatly appreciated! Best, Nancy Rusinak, Ed.S.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Re: parsing out one variable to many
|
|
In reply to this post by John F Hall
Nancy Forgot to say there is a set of syntax-based MULT RESPONSE tutorials on: http://surveyresearch.weebly.com/33-multiple-response-mult-response.html I noticed there was a set of 4 responses in one of your earlier mails to SPSS-X: I’ll work out something in SPSS syntax to spread responses out on to 15 columns. John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com Start page: www.surveyresearch.weebly.com/spss-without-tears.html |
Re: parsing out one variable to many
|
Administrator
|
In reply to this post by Bruce Weaver
Or
END LOOP IF (condition). condition: (#S=" ")? (#i=0) followed by slamming rest of #s into last vector position.
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
Re: parsing out one variable to many
|
Administrator
|
In reply to this post by John F Hall
@John,
"time = money and I don’t have time to learn about strings". Over the long haul you will save a lot of time=money by spending 20 minutes to learn about parsing strings. I am surprised/shocked/appalled to read this from someone who claims to have been using SPSS since 1972? --
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
Re: parsing out one variable to many
|
Administrator
|
In reply to this post by David Marso
Coup de Grace (Thread Killer: Aside from a vague feeling someone will toss some unnecessary python dependency into the mix). Shorter and a bit more general and eye friendly .
-- DATA LIST /strX (A80). BEGIN DATA 1,2,4,7 2,7,9,10,12 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 END DATA. DEFINE strSPLIT (Orig !TOKENS(1) /New !TOKENS(1) /NumX !TOKENS(1)/ Delim !TOKENS(1) !DEFAULT (",")). STRING #cpyx(A80). COMPUTE #cpyx=!Orig . VECTOR !New (!NumX). RECODE !CONCAT(!New,1," TO ", !New,!NumX, '(SYSMIS=0)'). FORMATS !CONCAT(!New,1," TO ", !New,!NumX, '(F1.0)'). LOOP. + COMPUTE #ind=INDEX(#cpyx,!QUOTE(!Delim)). + DO IF (#ind GT 0). + COMPUTE !New(NUMBER(RTRIM(SUBSTR(#cpyx,1,#ind-1)),F2))=1. + COMPUTE #cpyx=SUBSTR(#cpyx,#ind+1). + ELSE . + COMPUTE !New(NUMBER(RTRIM(#cpyx),F2))=1. + END IF. END LOOP IF #ind=0. !ENDDEFINE . strSPlit Orig=strX New=X NumX=15 . LIST. The variables are listed in the following order: LINE 1: STRX LINE 2: X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 STRX: 1,2,4,7 X1: 1 1 0 1 0 0 1 0 0 0 0 0 0 0 0 STRX: 2,7,9,10,12 X1: 0 1 0 0 0 0 1 0 1 1 0 1 0 0 0 STRX: 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 X1: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Number of cases read: 3 Number of cases listed: 3 ** Actual syntax generated by Macro **. Data List will read 1 records from the command file Variable Rec Start End Format STRX 1 1 80 A80 strSPlit Orig=strX New=X NumX=15 . 1107 M> 1108 M> . 1109 M> STRING #CPYX(A80). 1110 M> COMPUTE #CPYX= strX. 1111 M> VECTOR X ( 15 ). 1112 M> RECODE X1 TO X15(SYSMIS=0). 1113 M> FORMATS X1 TO X15(F1.0). 1114 M> LOOP. 1115 M> COMPUTE #IND=INDEX(#CPYX, ',' ). 1116 M> DO IF (#IND GT 0). 1117 M> COMPUTE X (NUMBER(RTRIM(SUBSTR(#CPYX,1,#IND-1)),F2))=1. 1118 M> COMPUTE #CPYX=SUBSTR(#CPYX,#IND+1). 1119 M> ELSE. 1120 M> COMPUTE X (NUMBER(RTRIM(#CPYX),F2))=1. 1121 M> END IF. 1122 M> END LOOP IF #IND=0. 1123 M> . ** Another version which merely spreads the values from left to right rather than dummy coding **. DATA LIST /strX (A80). BEGIN DATA 1,2,4,7 2,7,9,10,12 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 END DATA. DEFINE StrParse ( Origstr !TOKENS(1) / New !TOKENS(1) / FMT !TOKENS(1) !DEFAULT (A2) / NumElem !TOKENS(1) / Delim !TOKENS(1) !DEFAULT (",") ). STRING #cpyx(A80). COMPUTE #cpyx=!OrigStr . VECTOR !New (!Numelem,!FMT). COMPUTE #insert=1. !LET !expr = 'SUBSTR(#cpyx,1,#ind-1)'. !IF (!SUBSTR(FMT,1) !NE A) !THEN !LET !Expr=!CONCAT('NUMBER(RTRIM(SUBSTR(#cpyx,1,#ind-1)),',!fmt,')'). !IFEND LOOP. + COMPUTE #ind=INDEX(#cpyx,!QUOTE(!Delim)). + DO IF (#ind GT 0). + COMPUTE !New(#insert)=!Expr . + COMPUTE #cpyx=SUBSTR(#cpyx,#ind+1). + COMPUTE #insert=#insert+1. + ELSE . !LET !Expr= '#cpyx'. !IF (!SUBSTR(FMT,1) !NE A) !THEN !LET !Expr=!CONCAT('NUMBER(RTRIM(#cpyx),',!fmt,')'). !IFEND + COMPUTE !New(#insert)=!Expr . + END IF. END LOOP IF #ind=0. !ENDDEFINE . SET MPRINT ON PRINTBACK ON. strParse Origstr=strX New=X FMT=N2 NumElem=15 . LIST. The variables are listed in the following order: LINE 1: STRX LINE 2: X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 STRX: 1,2,4,7 X1: 01 02 04 07 . . . . . . . . . . . STRX: 2,7,9,10,12 X1: 02 07 09 10 12 . . . . . . . . . . STRX: 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 X1: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 Number of cases read: 3 Number of cases listed: 3
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
Re: parsing out one variable to many
|
|
The Python solution I posted yesterday
is much simpler, easier to understand, and easier to confirm correct.
Jon Peck (no "h") aka Kim Senior Software Engineer, IBM [hidden email] phone: 720-342-5621 From: David Marso <[hidden email]> To: [hidden email], Date: 04/23/2013 07:43 AM Subject: Re: [SPSSX-L] parsing out one variable to many Sent by: "SPSSX(r) Discussion" <[hidden email]> Coup de Grace (Thread Killer: Aside from a vague feeling someone will toss some unnecessary python dependency into the mix). Shorter and a bit more general and eye friendly . -- DATA LIST /strX (A80). BEGIN DATA 1,2,4,7 2,7,9,10,12 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 END DATA. DEFINE strSPLIT (Orig !TOKENS(1) /New !TOKENS(1) /NumX !TOKENS(1)/ Delim !TOKENS(1) !DEFAULT (",")). STRING #cpyx(A80). COMPUTE #cpyx=!Orig . VECTOR !New (!NumX). RECODE !CONCAT(!New,1," TO ", !New,!NumX, '(SYSMIS=0)'). FORMATS !CONCAT(!New,1," TO ", !New,!NumX, '(F1.0)'). LOOP. + COMPUTE #ind=INDEX(#cpyx,!QUOTE(!Delim)). + DO IF (#ind GT 0). + COMPUTE !New(NUMBER(RTRIM(SUBSTR(#cpyx,1,#ind-1)),F2))=1. + COMPUTE #cpyx=SUBSTR(#cpyx,#ind+1). + ELSE . + COMPUTE !New(NUMBER(RTRIM(#cpyx),F2))=1. + END IF. *END LOOP IF #ind=0.* !ENDDEFINE . strSPlit Orig=strX New=X NumX=15 . LIST. The variables are listed in the following order: LINE 1: STRX LINE 2: X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 STRX: 1,2,4,7 X1: 1 1 0 1 0 0 1 0 0 0 0 0 0 0 0 STRX: 2,7,9,10,12 X1: 0 1 0 0 0 0 1 0 1 1 0 1 0 0 0 STRX: 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 X1: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Number of cases read: 3 Number of cases listed: 3 ** Actual syntax generated by Macro **. Data List will read 1 records from the command file Variable Rec Start End Format STRX 1 1 80 A80 strSPlit Orig=strX New=X NumX=15 . 1107 M> 1108 M> . 1109 M> STRING #CPYX(A80). 1110 M> COMPUTE #CPYX= strX. 1111 M> VECTOR X ( 15 ). 1112 M> RECODE X1 TO X15(SYSMIS=0). 1113 M> FORMATS X1 TO X15(F1.0). 1114 M> LOOP. 1115 M> COMPUTE #IND=INDEX(#CPYX, ',' ). 1116 M> DO IF (#IND GT 0). 1117 M> COMPUTE X (NUMBER(RTRIM(SUBSTR(#CPYX,1,#IND-1)),F2))=1. 1118 M> COMPUTE #CPYX=SUBSTR(#CPYX,#IND+1). 1119 M> ELSE. 1120 M> COMPUTE X (NUMBER(RTRIM(#CPYX),F2))=1. 1121 M> END IF. 1122 M> END LOOP IF #IND=0. 1123 M> . ** Another version which merely spreads the values from left to right rather than dummy coding **. DATA LIST /strX (A80). BEGIN DATA 1,2,4,7 2,7,9,10,12 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 END DATA. DEFINE StrParse ( Origstr !TOKENS(1) / New !TOKENS(1) / FMT !TOKENS(1) !DEFAULT (A2) / NumElem !TOKENS(1) / Delim !TOKENS(1) !DEFAULT (",") ). STRING #cpyx(A80). COMPUTE #cpyx=!OrigStr . VECTOR !New (!Numelem,!FMT). COMPUTE #insert=1. !LET !expr = 'SUBSTR(#cpyx,1,#ind-1)'. !IF (!SUBSTR(FMT,1) !NE A) !THEN !LET !Expr=!CONCAT('NUMBER(RTRIM(SUBSTR(#cpyx,1,#ind-1)),',!fmt,')'). !IFEND LOOP. + COMPUTE #ind=INDEX(#cpyx,!QUOTE(!Delim)). + DO IF (#ind GT 0). + COMPUTE !New(#insert)=!Expr . + COMPUTE #cpyx=SUBSTR(#cpyx,#ind+1). + COMPUTE #insert=#insert+1. + ELSE . !LET !Expr= '#cpyx'. !IF (!SUBSTR(FMT,1) !NE A) !THEN !LET !Expr=!CONCAT('NUMBER(RTRIM(#cpyx),',!fmt,')'). !IFEND + COMPUTE !New(#insert)=!Expr . + END IF. *END LOOP IF #ind=0.* !ENDDEFINE . SET MPRINT ON PRINTBACK ON. strParse Origstr=strX New=X FMT=N2 NumElem=15 . LIST. The variables are listed in the following order: LINE 1: STRX LINE 2: X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 STRX: 1,2,4,7 X1: 01 02 04 07 . . . . . . . . . . . STRX: 2,7,9,10,12 X1: 02 07 09 10 12 . . . . . . . . . . STRX: 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 X1: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 Number of cases read: 3 Number of cases listed: 3 David Marso wrote > Or > END LOOP IF (condition). > condition: (#S=" ")? > (#i=0) followed by slamming rest of #s into last vector position. > Bruce Weaver wrote >> Gene, I / >> think / >> the problem was that your condition for exiting the loop was never >> becoming true, so it was only exiting the loop when it the loop number >> exceeded MXLOOPS. Given that we know 15 is the maximum number of loops >> (in this case, at least), if you LOOP just 15 times, it works with no >> warning messages. E.g., >> >> data list list / qID_9109(a36). >> begin data >> "1,2" >> "3,6,7" >> "2" >> "5,10,12,14" >> "1,2,3,4,5,6,7,8,9,10,11,12,13,14,15" >> end data. >> list. >> >> * Gene Maguin's code with some small modificiations. >> >> String #s(a36). >> Vector y(15,f2.0). >> Compute #s=qID_9109. >> If (substr(#s,char.len(#s),1) ne ',') #s=concat(rtrim(qID_9109),','). >> compute #k = 1. * >> Loop #k = 1 to 15. * >> Compute #i=index(#s,','). >> If (#i eq 2) y(number(substr(#s,1,1),f1.0))=1. >> If (#i eq 3) y(number(substr(#s,1,2),f2.0))=1. >> if (char.len(#s) ge 4) #s=substr(#s,#i+1). >> End loop. * >> recode Y1 to Y15 (sysmis=0). * >> LIST Y1 to Y15 qID_9109. >> >> OUTPUT: >> >> y1 y2 y3 y4 y5 y6 y7 y8 y9 y10 y11 y12 y13 y14 y15 qID_9109 >> >> 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1,2 >> 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 3,6,7 >> 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 2 >> 0 0 0 0 1 0 0 0 0 1 0 1 0 1 0 5,10,12,14 >> 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 >> 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 >> >> Number of cases read: 5 Number of cases listed: 5 >> >> >> Maguin, Eugene wrote >>> Here's a not so compact method. A key drawback is that when the loop >>> exits after processing a case a warning is issued. I don't know whether >>> there is a way to rewrite it to preclude the warning since you don't >>> know how many responses are concatenated unless you scoot through each >>> record and count the commas. Gene Maguin >>> >>> String #s(a36). >>> Vector y(15,f2.0). >>> Compute #s=qID_9109. >>> If (substr(#s,char.len(#s),1) ne ',') #s=concat(rtrim(qID_9109),','). >>> Loop if (char.len(#s) ge 2). >>> Compute #i=index(#s,','). >>> If (#i eq 2) y(number(substr(#s,1,1),f1.0))=1. >>> If (#i eq 3) y(number(substr(#s,1,2),f2.0))=1. >>> if (char.len(#s) ge 4) #s=substr(#s,#i+1). >>> End loop. >>> >>> >>> -----Original Message----- >>> From: SPSSX(r) Discussion [mailto: >>> SPSSX-L@.UGA >>> ] On Behalf Of Nancy Rusinak >>> Sent: Monday, April 22, 2013 9:54 AM >>> To: >>> SPSSX-L@.UGA >>> Subject: parsing out one variable to many >>> >>> I am collecting survey data. Respondents are asked to categorize their >>> comments into different buckets. They can choose as many as they like >>> or none at all. If they choose the first category, the data puts a "1" >>> in the field. If they choose the first and fifth category, the data >>> reads "1,5." >>> There are 15 possible choices. >>> >>> I'd like to create a variable called "strategy" for category 1 and have >>> a "1" put in that variable if the respondent chose that bucket; a >>> variable called "communication" for category 2 and a "1" put in that >>> variable if the respondent chose that bucket and so on. >>> >>> sample data, variable named qID_9109: >>> 1,2 >>> 3,6,7 >>> 2 >>> 5,10,12,14 >>> >>> Can anyone help here? >>> >>> ===================== >>> To manage your subscription to SPSSX-L, send a message to >>> LISTSERV@.UGA >>> (not to SPSSX-L), with no body text except the command. To leave the >>> list, send the command SIGNOFF SPSSX-L For a list of commands to manage >>> subscriptions, send the command INFO REFCARD >>> >>> ===================== >>> To manage your subscription to SPSSX-L, send a message to >>> LISTSERV@.UGA >>> (not to SPSSX-L), with no body text except the >>> command. To leave the list, send the command >>> SIGNOFF SPSSX-L >>> For a list of commands to manage subscriptions, send the command >>> INFO REFCARD ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/parsing-out-one-variable-to-many-tp5719621p5719658.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: parsing out one variable to many
|
Administrator
|
I don't see your post.
I have confirmed my code it correct. Feel free to try to break it! --
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
Re: parsing out one variable to many
|
|
In reply to this post by David Marso
David In all the years I've been designing and analysing surveys I've never needed to parse a script. I offered a solution to do the analysis in response to an off-list note from Nancy about MULT RESPONSE and copied my reply and her note to the list. In any case there were plenty of suggestions offered on parsing the script. The solution I offered took me about 20 minutes in which I modified Nancy's data, checked the FM, wrote and tested the syntax, then produced a specimen analysis. Nancy has never used MULT RESPONSE, but I'm sure she'll need it in her analysis. Your solution creates a string of 1s and 0s over 15 variables which can be analysed using MULT RESPONSE in dichotomous mode. My SPSS syntax deals with codes 1 - 15 in up to three variables, leaving the original values intact. A further solution would be for your code to be modified to yield codes 1 - 15 each with its own variable. It depends whether one wants to use MULT RESPONSE in integer or dichotomous mode. I've frequently used both (15 variable) conventions in the same survey on the same variables by recoding 1-15 to 1 or the series of 1s to 1 -15. I've dealt with coding conventions like these over many years, and run many an SPSS job to convert between one convention and another, but in the days of 80-column cards the fewer columns used the cheaper the data-prep. I'm busy writing tutorials, but will have a look at scripts etc when I have time. However, if they're not covered in Gradpack, I probably won't bother as I'm concerned mainly with entry level materials to get people started on, and getting an appetite for, the world of survey research and survey analysis. John John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com Start page: www.surveyresearch.weebly.com/spss-without-tears.html -----Original Message----- @John, "time = money and I don’t have time to learn about strings". Over the long haul you will save a lot of time=money by spending 20 minutes to learn about parsing strings. I am surprised/shocked/appalled to read this from someone who claims to have been using SPSS since 1972? -- John F Hall wrote > Nancy > > I can see your problem, but other listers have suggested solutions to > spread out the strings. I'm copying this back to the list so others > can see it. > > How many cases do you have? Are there only up to three responses for > each respondent? If so you need to spread the responses out on to > three separate columns, which is what you've tried to do with your > syntax. String manipulation is not my forte, I'm afraid. You can > apply the same logic if there are more than three responses per case. > > I'm always aware that time = money and I don't have time to learn > about strings. Provided the data set is not too large, I would > probably attempt something by hand such as copying the strings into > two more columns and then editing them. I would also try copying the > column for Q.66 into Word to edit into something that data list can > read as three variables (even if the 2nd and 3rd are blank). > > Here's my rough and ready solution: > > Copy your data into a txt file > > 0,1,14 > 2,6,7 > 10,11,15 > 3 > 0,4 > 8,10,12 > > > Replace all single digits with leading 0 and commas by blanks (I did > this by > hand) > > 00 01 14 > 02 06 07 > 10 11 15 > 03 > 00 04 > 08 10 12 > > . . and save to nancy.txt on f: > > SPSS syntax: > > data list file 'f:\nancy.txt' rec 1 > / x1 1-2 x2 4-5 x3 7-8. > > list /cases 6. > > freq var x1 x2 x3. > > mult resp groups q66 (x1 to x3 (0,15)) /freq q66. > > From: Nancy Rusinak [mailto: > nancy@ > ] > Sent: 23 April 2013 05:15 > To: John F Hall > Subject: Re: parsing out one variable to many > > Hello Mr. Hall, > > Thank you for writing & attempting to assist me. > > I'm afraid I have never worked with MULT RESPONSE syntax in the past so > it's > all Greek to me at this point. > > I have worked a good bit with SUBSTR and am trying to create substrings > using the "," as a delimiter. I'm almost there but, alas, not quite. Can > you > possibly help? > > Here's a sample of my data -variable is called Q66. It's string, 10 wide: > > 0,1,14 > 2,6,7 > 10,11,15 > 3 > 0,4 > 8,10,12 > > I only want 3 variables when I finish. For example, for the first row of > data, variable 1 would be "0," variable 2 would be "1," and variable 3 > would > be "14." > > Here's my syntax, which I found online and have been trying to adapt for > my > purposes: > > STRING #TEMP(A50). > COMPUTE #TEMP=Q66. > VECTOR VAR(3, A2). > LOOP #I=1 TO 3. > COMPUTE #INDEX=INDEX(#TEMP, ","). > COMPUTE VAR(#I)=SUBSTR(#TEMP,1, #INDEX-1). > COMPUTE #TEMP=SUBSTR(#TEMP, #INDEX+1). > END LOOP IF INDEX(#TEMP, ",")=0. > EXECUTE. > > When I run this, the first two variables are correctly populated for the > data strings that contain a "," but not for data strings that do not. > Those > are null. Also, the third variable is not populated for any of the data - > just null down the column. SPSS output logs the following: "the third > argument to SUBSTR (the length) is missing or otherwise invalid. The > argument must be a non-negative integer. The result has been set to the > null > string." This indicates to me that the #INDEX-1 and #INDEX+1 are an issue > but I do not know a workaround... > > Any help you could offer would be greatly appreciated! > > Best, > > Nancy Rusinak, Ed.S. > > Sent from my iPad > > On Apr 22, 2013, at 10:38 AM, John F Hall < > johnfhall@ > > wrote: > Nancy > > IN SPSS, easiest would be to have 15 variables, one for each response > category. If you label each value (1st variable only) you can then use > MULT > RESPONSE to tabulate the responses. > > val lab v1 1 'Strategy' 2 ' ~~~~' etc etc > mult resp category (v1 to v15 (1,15)) > /freq category. > > Another way is to label each variable and use MULT RESPONSE in dichotomous > mode, but to do this you first have to recode the values for all variables > to 1. I've played safe and kept your originals, but recoded to a new > variable > > recode v1 to v15 (1 thru 15 = 1)(else =0) into r1 to r15. > var lab r1 '~~~' > /r2 '~~~' > ~~~ > /r15 '~~~'. > mult resp rcategory (r1 to r15 (1)) > /freq rcategory. > > There are examples of this (illustrated by multiple response questions > from > the British Social Attitudes survey 1987 and the European Social Survey > 2002) in sections 3.2 and 3.3 (pp 27-35) and the 3rd slide show of my 2006 > Old Dog, Old Tricks presentation SPSS usage in major surveys. > (See: > http://surveyresearch.weebly.com/old-dog-old-tricks-using-spss-syntax-to-bea > t-the-mouse-trap.html) > > Get back to me if you don't understand that or if you need further help. > > > John F Hall (Mr) > [Retired academic survey researcher] > > Email: > johnfhall@ > Website: www.surveyresearch.weebly.com > Start page: www.surveyresearch.weebly.com/spss-without-tears.html > > > > > > > -----Original Message----- > From: SPSSX(r) Discussion [mailto: > ] On Behalf Of > Nancy Rusinak > Sent: 22 April 2013 15:54 > To: > Subject: parsing out one variable to many > > I am collecting survey data. Respondents are asked to categorize their > comments into different buckets. They can choose as many as they like or > none at all. If they choose the first category, the data puts a "1" in > the > field. If they choose the first and fifth category, the data reads "1,5." > There are 15 possible choices. > > I'd like to create a variable called "strategy" for category 1 and have a > "1" put in that variable if the respondent chose that bucket; a variable > called "communication" for category 2 and a "1" put in that variable if > the > respondent chose that bucket and so on. > > sample data, variable named qID_9109: > 1,2 > 3,6,7 > 2 > 5,10,12,14 > > Can anyone help here? > > ===================== > To manage your subscription to SPSSX-L, send a message to > <mailto: > > > (not to > SPSSX-L), with no body text except the command. To leave the list, send > the > command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, > send > the command INFO REFCARD ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/parsing-out-one-variable-to-many-tp5719621p5719656.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
| Free forum by Nabble | Edit this page |